Looking at the TCP/IP stack, it seems that you are not allow to bypass any of the layers. For instance, you can't go from application layer straight to layer 3 (network).
So, let's say we have two routers connected to each other. Would you describe it as a Layer 3 infrastracture? Or would you also call it a Layer 2?
I'm saying Layer 2 also since the IP packets have to be encapsulated to a Layer 2 frame before it gets sent out to the other router via a, for example, Ethernet interface.
The wise Ron J enlightened me with:
Then I'd suggest you call it a layer 1 connection, since that layer2 frame is also an abstraction. Just like the layer3 etc
Mark
Routers primarily operate at layer 3, but to get stuff out an ethernet interface, you will always have to move down from layer 3 to 2 to 1, and on the other side back from 1 to 2 to 3.
They're *not* connected from layer 3 on router 1 to layer 3 on router 2 -- they're connected on router1 from layer 3 to 2 to 1 to layer 1 on router 2 to layer 2 to layer 3. Each device has a full 7 layer TCP/IP stack, but primary operation for a router (ie, anything that's got IP addresses that aren't the router itself) is to take the bits, make them into frames, strip ethernet headers into packets, determine the outgoing interface for that packet, reencapsulate into ethernet frame, and down to layer 1 again.
In a switch, most data (anything not intended for the switch itself) won't move up its stack beyond layer 2 before exiting downward again.
Jasper
Have something to add? Share your thoughts — no account required.
Ask the community — no account required